DNA and RNA polymerases and the technologies they enable have revolutionized biomedical research. However, the exquisite specificity of natural polymerases limits their potential applications to those involving the fully natural biopolymers, which are unsuitable for many diagnostic, therapeutic, and DNA sequencing applications. The first two applications are of obvious relevance to human health, promising to revolutionize disease detection and treatment, and the latter promising to usher in an unprecedented era of personalized medicine. To address these limitations, we developed an activity-based selection system to evolve polymerases that recognize modified substrates. The system is based on co-display of polymerase libraries and substrates on bacteriophage particles and which allows for their diversification and selection for unnatural activities in a manner that imitates Darwinian evolution in nature. While we have identified several aspects of the system that still require optimization, we have already used it to evolve several first generation unnatural DNA polymerases that possess increased abilities to synthesize polymers comprised of nucleotides modified for different applications. For example, SFM19 is able to efficiently synthesize short stretches of polymers comprised of C2'-OMe modified nucleotides, which have potential applications as biostable polymers for diagnostic and therapeutic applications. Sf197 is able to more efficiently polymerize nucleotides modified for labeling and next-generation sequencing applications. While both evolved polymerase represent important first steps toward practically useful enzymes, they both still require further optimization: SFM19 for the synthesis of longer modified polymers, and Sf197 for increased efficiency. Our first objective is to further optimize our selection system and to adapt it for the evolution of RNA polymerases. Our second objective is to evolve polymerases with real, practical utility. As part of our second objective, SFM19 and Sf197 will each be further diversified and subjected to selections for optimized activity. We will also evolve an RNA polymerase to efficiently recognize C2'-OMe nucleotides and a DNA polymerase that enables the direct sequencing of methylated cytosines, which are central epigenetic markers whose distribution through the genome has critical health implications, but which is currently challenging to characterize. Achieving these objectives will deliver a robust system for evolving polymerases with specifically tailored activities, and four evolved polymerases that have immediate and important health related applications. Perhaps most importantly, the proposed research should illustrate the potential of polymerase evolution and reduce it to a more practical and user friendly system, with the goal of providing to the broader research community a generally accessible method to tailor polymerases for as many different activities as there are potential applications.